ESG and artificial intelligence
One of the most pressing issues is the lack of reliable and consistent ESG data. The lack of ESG data standards is a major barrier to increased adoption of sustainable investing. As a result, investors are forced to interpret unstructured data themselves, which hinders investment and discredit.
Artificial intelligence (AI) is a better way to address this data challenge.
Therefore, in this article, we will answer the following questions: How can AI optimize ESG investing today?
Several environmental, social and governance themes emerged after the Rio Earth Summit in 1992,such as impacts on biodiversity, respect for human rights, business ethics and of course climate change and the Covid-19 pandemic. Accurately measuring the economic and financial impacts associated with these topics requires access to relevant data, which can sometimes be difficult.
Indeed, the available information is often insufficient and heterogeneous. Therefore, the collection and use of this data is a complex and important process that can be greatly facilitated by the appropriate use of artificial intelligence (AI) derived technologies.
Definition of Artificial Intelligence (AI)
Artificial intelligence (AI) is a generic term that covers a variety of methods and technologies that allow machines to perform tasks that are generally in the realm of human intelligence. This means that machines can perform automated tasks (as in finance, “exclude securities from investment as long as the debt/EBITDA ratio is greater than 6”), but also support more complex functions such as voice or facial recognition, the development of autonomous cars, or in finance, “decide to keep the
securities in the portfolio despite the ratio doubtful accounts/total operating income if that makes sense”, using the previous example. Artificial intelligence brings together various techniques, including “Machine Learning” (or “machine learning process”), which is particularly important in asset management.
Artificial intelligence is not as new as people think. Work on this topic began in the late 1950s. But today, its use has become a reality due to the abundance of available and accessible data (“open data”), the ability to collect information has been strengthened by the development of increasingly powerful and faster processors, low-cost storage capacity Exponential growth, finally thanks to the
availability of “open source” software (the code of which is freely available).
The “Machine Learning” principle:
Machine learning is a subset of artificial intelligence and a branch of computing and statistics. It relies on algorithms, computer programs that process large amounts of data, to enable machines and/or systems to “learn” through experience in an automated way to improve their problem-solving abilities without humans having to explicitly program to do so.
AI’s contributions in the field of sustainable development and ESG
They are multiple, but perhaps one of the most relevant uses is to “fill in the gaps”, which is to find or build missing data. For example, it is sometimes difficult to obtain detailed information on small or unlisted companies, reducing the obligations to disclose or publish ESG data. On the other hand, these companies generally only communicate in the local language and meet unique local obligations and standards. This is the case in China, where asset managers often face little or no information about these companies. To avoid these difficulties, “machine learning” techniques can be used, using search algorithms, missing, or standardizing information can be collected or completed. These algorithms can be implemented by traditional rating agencies or ESG research providers, or even by fintech.
On the other hand, one of the main benefits of AI technologies is to provide management with a truly tailored approach. By adjusting algorithms, proprietary rating systems can be built by finding information deemed most relevant and therefore most likely to help generate financial performance (called “alpha”) or reduce risk levels. Marketplace (“Beta”). Within the framework of the ESG investment approach, where the key issues identified by the economic sectors were previously defined internally, AI can contribute to the reading grid by accessing information from many sources, most of which are freely accessible, especially on the Internet, thus enriching the database in real time.
The main contribution of AI in terms of investment decisions
Ideally, AI will lead to smarter management decisions. The idea is to develop an ESG approach that is more motivated by seeking investment opportunities than simply identifying risks. Indeed, the positive stock market performance of companies is often correlated with their excellence in ESG
strategy, but especially with their progress against these criteria (what we call “ESG momentum” or “ESG dynamics”). From AI will have the potential to outperform current approaches that are primarily based on ESG risk.
As various AI technologies can identify economic and financial opportunities and risks more quickly, this should also improve the accuracy and timeliness of investment decisions, increasing overall efficiency. This means that the internal organisation allows rapid integration of this information into the decision-making processes of managers. More accurate identification of financial opportunities
and risks should also improve the quality of management decisions, often when using information that is not explicitly disclosed by the company, or when public information may contain bias. Finally, a broader and more accurate identification of financial opportunities and risks should in fact lead to broader investment horizons.
AI capabilities and ESG specificities
It’s true that AI can definitely help, but it can’t do everything either. This is a very useful complement, but other elements are also needed, such as the definition of international ESG standards, which are still in their infancy.
For example, social factors are the most difficult to collect geographically comparable data. AI should help to assess the social pillars of the GSS, because even on basic metrics such as workplace accidents and their likelihood of occurrence, the data that companies report is very heterogeneous.
By applying its models, AI can reconcile data across geographic regions.